The Research On Intelligent Control Of Multi-reactors Nuclear Power Plant

PWR-type nuclear power plant with multi-module integrated reactors has the advantages of better economics of mass production construction,higher inherent safety and stability,and better space optimization.However,the complex coupling dynamics caused by the parallel operation of multiple reactor units put forward higher requirements for control system.Therefore,it is necessary and valuable to study the intelligent control method and controller of multi-reactor nuclear power plant.Firstly,I designed the main structure and parameters of the multi-stack nuclear power unit based on IP200,and established a real-time transient analysis model using the RELAP5 program according to the characteristics of the system structure.Then,the rationality and accuracy of the established model were verified by comparing the steady-state parameters of the system and the transient characteristics of the rising and falling power systems.According to the operating characteristics of the nuclear power plant of the sleeve-type DC evaporator reactor,the control strategies including the operation mode of the reactor load following power,the single reactor dual constant operation scheme,and the load distribution scheme under different operating conditions were designed.Based on the classic PID control principle,MATLAB was used to establish a multi-reactor nuclear power unit control system including reactor power controller,DC evaporator feed water controller and regulator pressure controller.Based on the classic PID control principle and MATLAB software,a multi-reactor nuclear power plant control system including a reactor power controller,a DC evaporator feed water controller,and a regulator pressure controller was established.Based on this control model,combined with existing control strategies,the intelligent control algorithms such as neural network control,genetic parameter tuning algorithm,adaptive dynamic programming control,and deep adaptive dynamic programming control were used to optimize the control system.On the Sim Exec platform,integrated joint debugging and real-time simulation with real-time RELAP5 program were achieved by UDP communication method.Through the simulation calculation of the working conditions such as sharp rise,load reduction and parallel variable load with unequal power,the feasibility of the intelligent control algorithm applied to the control of multi-stack nuclear power system and the advantages and disadvantages of the regulation performance of each intelligent controller were verified.The analysis of the results of various operating conditions shows that the genetic parameter tuning operation can significantly improve the adjustment performance of the neural network controller,and the ADP and DADP controllers after excellent data training can achieve better expected control results.All three intelligent controllers can meet the control needs of multi-core nuclear power units under expected operating conditions and their performance is better than PID control.Therefore,the control strategy and corresponding intelligent controller designed in this paper can realize more flexible and stable operation of multi-stack nuclear power units,improve the system’s operating characteristics and parameter dynamic quality.The research results in this paper can provide a certain theoretical basis for the intelligent control research of multi-reactors nuclear power plant.